You just bought Literotica...what do you change?

[MediocreAuthor]

As for Anal, it generally can fit into one of three categories, 1) gay, obviously, 2)pegging-fetish or LW. 3) LW obviously anal or dp. I only peruse that category for categories 2 or 3. so yes, get rid of Anal and force authors to post to one of the other three categories

I don't often read anal, but the first one that comes to mind I found on the story feedback forum, and it doesn't fit any of the three you listed.

It's basically EC but they have butt-sex.

 

[burgwad]

That is a very big "if". I've already talked about why it could easily go the other way and make story scores more vulnerable, both to intentional manipulation and to volatility.

I appreciate the lengthy response, but I'm not 100% clear on where we're disagreeing. If you have a sec, please feel free to hook me up with a quote/link to the comment where you addressed ("already talked about") this stuff. I did go looking myself, but your contributions to this thread are many, and I wasn't sure which specific one you had in mind.

 

[Bramblethorn]

I appreciate the lengthy response, but I'm not 100% clear on where we're disagreeing. If you have a sec, please feel free to hook me up with a quote/link to the comment where you addressed ("already talked about") this stuff. I did go looking myself, but your contributions to this thread are many, and I wasn't sure which specific one you had in mind.

Sure. I've discussed it in more detail in older threads on the same topic, but the key part is this:

I say this every time, but: removing anon voting would make vote-rigging worse, not better. The fewer the votes on each story, the fewer votes a cheat has to cast in order to change the outcome of a contest/etc., and the harder it is to identify that cheating.

I have an obnoxious migraine just now and it's less effort to retype stuff than to find those older posts, so in a bit more depth:

For the sake of example, let's suppose I've written a story that has an average of 4.90 off 150 votes, and somebody wants to knock it down to 4.85 to get it off the category front page or out of a competition or whatever.

To do that, they need to cast two one-star votes. (The calculation is: 150 * (4.90 - 4.85)/(4.85 - 1.00).)

If Laurel goes looking through the voting, two one-star votes in quick succession are going to look pretty suspicious, especially if technical evidence associated with those votes suggests they're coming from the same person. They're probably going to get swept.

Now consider the same scenario in a world where we cracked down on anon voting, which would inevitably result in some people no longer voting at all rather than registering/enabling cookies/whatever.

If, say, half the people who used to vote no longer do, then my story now has 4.90 off 75 votes, and it only takes a single one-star vote to knock it down to 4.85.

A single one-star vote on a story that was scoring 4.90? Might arouse suspicions, but there's no way to look at that single vote and determine that it's illegitimate. Maybe somebody read my story, genuinely hated it, and voted as the site permits them to do.

If you plug in different assumptions, of course you'll get different numbers out. But in every case, the more votes a story has, the less impact one downvote has on its score - which increases the number of votes somebody needs to cast to do their desired amount of damage, which makes them easier to catch.

Now let's consider the happy scenario where nobody is bombing stories at all. What happens then?

Anybody who's watched their scores when a new story goes live will know that they're very volatile at first, and then as more votes come in they tend to stabilise. When a story only has a few votes, its score is very susceptible not just to deliberate sabotage but also to random chance. Maybe your story posts on a day when more grouchy people just happen to be around and reading the category; maybe mine posted on a day when more nice people were around.

For stories in the high 4s, which is where people tend to worry the most about this kind of stuff, the random "noise" in the score is on the order of:

2*sqrt((5-S)/V)

where S is the score, and V is the number of votes.

(Math nerds: estimating two standard deviations, assuming binomial distribution dominated by 4s and 5s, usual independence assumptions, applying a couple of fudges as this isn't meant to be very precise.)

For my hypothetical story with 4.90 off 150 votes, the noise in that score is about +/- 0.05. Comparing that to somebody else's story with 4.85, the difference between my score and theirs isn't statistically meaningful - it's small enough that it could all be just down to luck. If I'd posted this on a different day, with different luck, it could've scored a 4.95 instead. In some categories the difference between 4.95 and 4.85 is the difference between #1 place on the all-time toplist, and not making the top 250.

If you halve the number of people voting, that +/- 0.05 becomes more like +/- 0.07, and the scores have become even less meaningful.

In many categories, I'd hazard a guess that the race between the top few places in things like the monthly competitions is decided more by that random noise than by any deliberate manipulation or by the relative merits of the top contenders. Making it harder for people to vote means randomness becomes even more of a factor.

If we're happy to accept a high level of randomness as the price for eliminating cheating, we could do it by ditching voting altogether and drawing contest winners out of a hat. But I think what most people want is a system where author skill/story quality counts for as much as possible. If that is what we want, then it's worth understanding that the noise can be a bigger problem than the cheats.

 

[MelissaBaby]

Why do people always assume that one bombs are cast by anonymous?

 

[MediocreAuthor]

Why do people always assume that one bombs are cast by anonymous?

I always assumed that illegitimate one bombs are generally anonymous.

If a member hates a story, and wants to vote it one star, that should be allowed (the one star button is there for a reason)

But in the context of vote rigging, the assumption is that people might down vote a story several times anonymously, rather than go through the trouble of creating multiple accounts.

That said, I suppose there's nothing preventing cheaters from making multiple accounts, but if Laurel can see that all an account has done has been downvote every story in a contest and upvote a single story... That's clear sign of tampering, and she can just disregard all of that accounts votes from that account going forward (I assume)

 

[TxRad]

I wouldn't buy it. I'd start a new place from scratch where the authors were anonymous and the readers had to log in.

 

[MelissaBaby]

I always assumed that illegitimate one bombs are generally anonymous.

If a member hates a story, and wants to vote it one star, that should be allowed (the one star button is there for a reason)

But in the context of vote rigging, the assumption is that people might down vote a story several times anonymously, rather than go through the trouble of creating multiple accounts.

That said, I suppose there's nothing preventing cheaters from making multiple accounts, but if Laurel can see that all an account has done has been downvote every story in a contest and upvote a single story... That's clear sign of tampering, and she can just disregard all of that accounts votes from that account going forward (I assume)

I suspect that it’s not uncommon for authors with a story that has a 4.94 rating to decide that all the stories in the category with ratings above that suck and deserve ones.

 

[AwkwardMD]

I suspect that it’s not uncommon for authors with a story that has a 4.94 rating to decide that all the stories in the category with ratings above that suck and deserve ones.

Whenever I get a monthly contest win, it's a surprise. The story in question will be floating around with a score in the high 4.7s or low 4.8s, but leaps up into the 4.9s with a sweep.

Within a week, it's back down to 4.8.

This has happened to me multiple times, and it has forced me to adjust my expectations for "what a good reception looks like."

 

[AwkwardMD]

What would it look like if you weighted the most popular vote? Like, taking Bramblethorn's above 4.9 story that gets down voted maliciously. If the overwhelming majority of votes are 5s, is there a downside to making them worth proportionately more?

 

[AwkwardlySet]

I'd say that the whole point of being able to disallow anonymous voting is that every vote cast would then be linked to a specific account. As a moderator, you can now see who is casting dubious votes and act accordingly. Sure, there are alt accounts all over Lit, but those can get banned easily, and applying for a new one takes about half a day or so? As an anonymous, your vote is tied to an IP (I assume) so all you need to do is use some kind of VPN to be able to vote again and again. By allowing authors to disable anonymous voting, the total amount of votes would be lower, that is true, and that fact would make the rating more vulnerable, but the benefits of being able to track all the votes and the time and effort it takes to create a new Lit account, (Creating a new email and then applying and being approved etc.) would make contests more easily controlled than they are now.

 

[MelissaBaby]

Whenever I get a monthly contest win, it's a surprise. The story in question will be floating around with a score in the high 4.7s or low 4.8s, but leaps up into the 4.9s with a sweep.

Within a week, it's back down to 4.8.

This has happened to me multiple times, and it has forced me to adjust my expectations for "what a good reception looks like."

Queen of the Roller Derby had a score of 4.98 with 99 votes. I kept refreshing the browser until it hit 100, so I could screenshot the amazing occurrence of having the #1 spot on the All Time Top List.

15 minutes later, with something like 104 or 105 votes, it was at 4.92.

It’s a 14 page story. I don’t think I’m paranoid to believe some of those four or five votes don’t reflect someone’s sincere belief in the quality of the story.

 

[AwkwardMD]

Queen of the Roller Derby had a score of 4.98 with 99 votes. I kept refreshing the browser until it hit 100, so I could screenshot the amazing occurrence of having the #1 spot on the All Time Top List.

15 minutes later, with something like 104 or 105 votes, it was at 4.92.

It’s a 14 page story. I don’t think I’m paranoid to believe some of those four or five votes don’t reflect someone’s sincere belief in the quality of the story.

4.98 with 99 votes is an absurdly impressive streak of votes. Kudos.

 

[MediocreAuthor]

I suspect that it’s not uncommon for authors with a story that has a 4.94 rating to decide that all the stories in the category with ratings above that suck and deserve ones.

That is an excellent point. That said, I don't think that it would be right to totally ban those votes.

Obviously, such voting is selfish, petty, self-centered, but outright banning it would also be unfair. I have seen contests in which I thought that the winning entry was bad... straight up trash.

I didn't uno-bomb the story, even though I hated it, but if I had uno-bombed it, it wouldn't be right to remove my vote. People are entitled to differing opinions.

 

[33RPMVinylPants]

Why not hide the score on contest submissions until the end of the contest?

If you don't know who's winning, you can't 1-star them to make yours rise in the ranks.

 

[MillieDynamite]

Who would do that shit? What does it benefit them? They don't earn anything from doing that but false pride in their highly questionable talent. I mean, if you have to cheat, how fucking good can you really be?

I suspect that it’s not uncommon for authors with a story that has a 4.94 rating to decide that all the stories in the category with ratings above that suck and deserve ones.

 

[AwkwardMD]

Who would do that shit? What does it benefit them? They don't earn anything from doing that but false pride in their highly questionable talent. I mean, if you have to cheat, how fucking good can you really be?

Some people go through life projecting this attitude onto others, which is how they justify doing it to others. They assume, then, that their score has had the same thing done to them.

Serial cheaters are often quick to accuse others of cheating, and use that same justification to cheat. Projection is a basic human behavior.

 

[Bramblethorn]

What would it look like if you weighted the most popular vote? Like, taking Bramblethorn's above 4.9 story that gets down voted maliciously. If the overwhelming majority of votes are 5s, is there a downside to making them worth proportionately more?

Currently, scores are calculated as an unweighted average:

(sum of all votes)/(total number of votes)

This is equivalent to

(1*x1 + 2*x2 + 3*x3 + 4*x4 + 5*x5)/(x1 + x2 + x3 + x4 + x5)

where x1 is the number of 1s, x2 is the number of 2s, and so on.

I think what you're proposing is to weight each individual vote according to how common that particular vote is? That would work out to:

(1*x1^2 + 2*x2^2 + ... + 5*x5^2)/(x1^2 + x2^2 + ... + x5^2)

Under this method, for a story that's mostly scoring high, the impact of the first few 1* votes would be very small, which might be good. For instance, if a story has 90 4s (score = 4.0) and then somebody hits it with ten 1s, the updated score is:

(1*10^2 + 4*90^2)/(10^2 + 90^2) = 3.96

But it creates some weird situations where lower votes can be better than higher ones. For instance, say I have a story that's got 30 1* and 70 4* votes. Score is:

(1*30^2 + 4*70^2)/(30^2 + 70^2) = 3.53.

But what if we upgrade twenty of those 4* votes to 5*?

(1*30^2 + 4*50^2 + 5*20^2)/(30^2 + 50^2 + 20^2) = 3.39.

Increasing those votes from 4* to 5* actually lowered the story score - splitting between 4s and 5s reduces the overall weight from those high votes, giving the 1s more influence, and in this scenario that matters more than the improvement from 4 to 5.

There might be fancier ways to do it that avoid that problem. But I do see complexity as a down-side in itself - the harder it is for people to understand the scoring system, or why it's producing results they didn't expect, the more potential there is for drama.

A slightly different option that gets suggested here occasionally is to ignore, say, the top and bottom 10% of votes and calculate the average from the rest, like in gymnastics scoring where the highest and lowest judges are ignored. This avoids some of those problems, but it loses discrimination at the top end - if we trim the top/bottom 10%, all the stories currently scoring 4.90 or above would have perfect 5.0s and there'd be a sea of "equal first".

I'd say that the whole point of being able to disallow anonymous voting is that every vote cast would then be linked to a specific account. As a moderator, you can now see who is casting dubious votes and act accordingly. Sure, there are alt accounts all over Lit, but those can get banned easily, and applying for a new one takes about half a day or so? As an anonymous, your vote is tied to an IP (I assume) so all you need to do is use some kind of VPN to be able to vote again and again. By allowing authors to disable anonymous voting, the total amount of votes would be lower, that is true, and that fact would make the rating more vulnerable, but the benefits of being able to track all the votes and the time and effort it takes to create a new Lit account, (Creating a new email and then applying and being approved etc.) would make contests more easily controlled than they are now.

I think sweeps make use of other technical information beyond IP address to identify malicious votes. (Not certain, might be misremembering.) But Laurel has asked that people not discuss the technical aspects of how sweeps work/how they might be defeated, and threads have been locked for that reason in the past, so we probably shouldn't get any more specific.

 

[M_K_Babalon]

There should just be an option to turn off anon voting, like comments. I suspect few who hate anon would take that risk, or quickly change their mind depending how their future stories voting turns out.

 

[MelissaBaby]

4.98 with 99 votes is an absurdly impressive streak of votes. Kudos.

Thank you. It was insane. And very gratifying, because I felt like I really stretched myself on that story.

 

[MelissaBaby]

That is an excellent point. That said, I don't think that it would be right to totally ban those votes.

Obviously, such voting is selfish, petty, self-centered, but outright banning it would also be unfair. I have seen contests in which I thought that the winning entry was bad... straight up trash.

I didn't uno-bomb the story, even though I hated it, but if I had uno-bombed it, it wouldn't be right to remove my vote. People are entitled to differing opinions.

I'm content to trust to sweeps to deal with the problem. And besides, I would look like a real ass pitching a bitch about the scoring on a story with a 4.91. I only presented it as an example because I watched the scoring change in real time.

 

[AwkwardlySet]

I understand the urge to keep the exact technical details of the way voting and sweeps work away from the public. But some things I can deduce from my own experience. Back when I was being targeted by the bombing campaign of certain individuals, I took care to observe how the bombing occurred. Since the stats in the control panel of an author refresh immediately, I noticed a very distinct pattern.

Every 4-5 hours or so, my stories would get six or seven 1* votes (on each story) all cast together more or less. It makes sense to assume that it was the same small group of individuals who would cast the votes and then after several hours, they would be able to do it once again. Maybe they used VPNs, maybe they used something else, I have no idea. For some reason, they would need to wait several hours to cast those votes again, VPN or not.

There is also an alternative explanation. It was all done by the same person who used the VPN or some other method to repeatedly change their IP and cast one 1* after another until they got bored. Clicking on each of my 20-ish stories and then clicking on the last page and then casting the 1* vote, and then changing the IP and then repeating the whole deal can get tiresome, so it makes sense that that person would get bored after casting several bombs on each story. But then after a couple of hours, they would get motivated to do it once again. Since I am inclined to believe that it was actually done by just one person, or maybe two, rather than a group, I would say that maybe they would decide to repeat the process once my author's page (the way readers see it) refreshed. Seeing my scores going down could possibly prompt them to repeat the process.

The first explanation would imply that Lit has some kind of voting abuse protection that doesn't extend past a couple of hours, so it is not much of a protection after all.
The second implies that there is no protection at all and that presumably, a simple IP change can make one able to cast votes repeatedly. There are likely other explanations, but these are the ones that make the most sense to me, based on the pattern I was able to observe over a period of two weeks or so before I decided to take my stories down.

 

[BobbyLaker]

I'd say that the whole point of being able to disallow anonymous voting is that every vote cast would then be linked to a specific account. As a moderator, you can now see who is casting dubious votes and act accordingly. Sure, there are alt accounts all over Lit, but those can get banned easily, and applying for a new one takes about half a day or so? As an anonymous, your vote is tied to an IP (I assume) so all you need to do is use some kind of VPN to be able to vote again and again. By allowing authors to disable anonymous voting, the total amount of votes would be lower, that is true, and that fact would make the rating more vulnerable, but the benefits of being able to track all the votes and the time and effort it takes to create a new Lit account, (Creating a new email and then applying and being approved etc.) would make contests more easily controlled than they are now.

... and your home IP tends to change now and again anyway, at least in the UK.

 

[MelissaBaby]

If people want to bomb, they are going to bomb. What's to stop someone from voting on their home computer, logging into McDonald's wifi on their phone and voting again, then voting a third time on their work computer?

 

[MillieDynamite]

I don't know, ethics? I guess that's too much for us to expect, though.

If people want to bomb, they are going to bomb. What's to stop someone from voting on their home computer, logging into McDonald's wifi on their phone and voting again, then voting a third time on their work computer?

 

[AwkwardlySet]

If people want to bomb, they are going to bomb. What's to stop someone from voting on their home computer, logging into McDonald's wifi on their phone and voting again, then voting a third time on their work computer?

Well, yeah, exactly... What would stop them is the author being able to turn off anonymous voting. Then, it doesn't matter where you are, or what computer you are using. Every vote would be tied to an actual account. The solution is not perfect, sure, but I believe it is better than what we have now.